netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Johannes Berg <johannes@sipsolutions.net>
To: Dan Williams <dcbw@redhat.com>
Cc: netdev <netdev@vger.kernel.org>, Jiri Benc <jbenc@suse.cz>,
	"John W. Linville" <linville@tuxdriver.com>,
	Larry Finger <Larry.Finger@lwfinger.net>
Subject: Re: [RFC] cfg80211 and nl80211
Date: Thu, 05 Oct 2006 09:47:32 +0200	[thread overview]
Message-ID: <1160034452.2728.2.camel@ux156> (raw)
In-Reply-To: <1159805711.2834.67.camel@localhost.localdomain>

Umm, looks like I skipped this paragraph in my earlier reply to you. Sorry
about that.

> I'd also argue that one specific BSSID is part of an initial
> configuration.  We should support that in config command.  It's an
> implicit SET_FIXED_BSSID, yes.  But one of the major points of
> nl80211/cfg80211 was that you could bundle up a set of configuration
> settings into a single atomic "packet", which you couldn't do with WE.
> 
> So if a specific BSSID isn't sent in the initial config command, when do
> you set a specific BSSID?  Before?  After?  The behavior starts getting
> complicated, and we're back to a situation where every driver implements
> the semantics in a slightly different manner.

Ah, good point. But then, why would you want to set a specific (initial)
BSSID at all? Either you set userspace roaming (which you'd do before
setting the SSID) then the kernel can't do anything without you setting a
BSSID, or you don't set userspace roaming, then all the kernel needs is the
SSID.

I'm thinking you probably want something like 'list of BSSIDs to use for
userspace roaming' and possibly a blacklist too, although I'm inclined to
let userspace manage the blacklist by way of having a whitelist *only* and
having userspace simply add everything to the whitelist that it discovers
through scanning and isn't on the blacklist...

Hence, would you be satisfied with a BSSID-whitelist for kernel-controlled
roaming (userspace roaming doesn't need the kernel to know about the
whitelist)? Heck, you could even use a single-element whitelist for when you
want to force the kernel to associate to that AP... Maybe we should thus
drop the userspace roaming support? I think it's a simpler API though...
Then again, why do we need a BSSID-whitelist? Just have userspace control
roaming then...

Also, the use case you want could probably be achieved by turning on
userspace roaming, setting the BSSID for it, configuring the SSID and then
turning off userspace roaming again.

Or let me put it another way: I'm not sure what the use case actually is :)

johannes

      parent reply	other threads:[~2006-10-05  7:46 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-09-28  9:23 [RFC] cfg80211 and nl80211 Johannes Berg
2006-09-29 21:10 ` James Ketrenos
2006-09-30  3:00   ` Michael Wu
2006-10-02  9:08   ` Johannes Berg
2006-09-30  3:14 ` Michael Wu
2006-10-02 16:15 ` Dan Williams
2006-10-02 17:01   ` Dan Williams
2006-10-04  7:41   ` Johannes Berg
2006-10-04 14:19     ` Johannes Berg
2006-10-04 17:57       ` Dan Williams
2006-10-05  7:59         ` Johannes Berg
2006-10-05 13:13         ` Stuffed Crust
2006-10-05 15:46           ` Jouni Malinen
2006-10-05 16:20         ` Jouni Malinen
2006-10-06  9:41           ` Johannes Berg
2006-10-05  7:47   ` Johannes Berg [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1160034452.2728.2.camel@ux156 \
    --to=johannes@sipsolutions.net \
    --cc=Larry.Finger@lwfinger.net \
    --cc=dcbw@redhat.com \
    --cc=jbenc@suse.cz \
    --cc=linville@tuxdriver.com \
    --cc=netdev@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).